My Journey in Nielsen

• Authoring
  
• My Daily Routine
• VBA journey
• DCT
• Support team
• Finding ad hoc
• Analysis through R

• Elegant, flexible, and fast dynamic report generation with R:

• https://yihui.name/knitr/options/#text-results -

• Compiling and tools

• R Studio auto formats and runs the code when you save the document

• How to download R and create graph: http://rpubs.com/PeterDat/341231

• Listening on Google: https://goo.gl/bdFLGG Many interesting things
• Team drive: R training Vietnam: https://goo.gl/K8BeN4 Easy to learn and reproducible

• 1) Going on Fields
• 2) Collection data on paper
• 3) Why not work on DCT?

We could use lookup function :

• To help apply store AQCC planning For DA
• To help FA to reduce fill store's info
• Task: Teams manage Attributes for Store

when looping Photo, the memory limits in the FA's phone.

• Step 1: Mr.An gives me raw and excel pilot VPM analysis
• Step 2: Ms.Loi gives me the pptx to insert data
• Step 3: I am using R to extract data

Benefit: double check with excel pilot VPM In the future can connect with Tableau to create more beautiful chart.

library(ggplot2)
library(readxl)

NSO1 <- read_excel("NSO_Dashboard_Trend_Wave4_2018-01-17.xlsx", sheet = "Data_Source", 
                                                   range = "A6:GS15321")

store <- aggregate(  `Respond ID` ~ City+ Timecode + Channel ,data= NSO1,length)

storehcm <- store[store$City == "HO CHI MINH",]
storehcm[storehcm$Timecode == "Wave 1",]$`Respond ID` <- (storehcm[storehcm$Timecode == "Wave 1",]$`Respond ID`)/sum(storehcm[storehcm$Timecode == "Wave 1",]$`Respond ID`)
storehcm[storehcm$Timecode == "Wave 2",]$`Respond ID` <- (storehcm[storehcm$Timecode == "Wave 2",]$`Respond ID`)/sum(storehcm[storehcm$Timecode == "Wave 2",]$`Respond ID`)
storehcm[storehcm$Timecode == "Wave 3",]$`Respond ID` <- (storehcm[storehcm$Timecode == "Wave 3",]$`Respond ID`)/sum(storehcm[storehcm$Timecode == "Wave 3",]$`Respond ID`)
storehcm[storehcm$Timecode == "Wave 4",]$`Respond ID` <- (storehcm[storehcm$Timecode == "Wave 4",]$`Respond ID`)/sum(storehcm[storehcm$Timecode == "Wave 4",]$`Respond ID`)

hcmstore<- setNames(storehcm,  c( "City" ,"Timecode","Channel","% store handling"))


ggplot(data=storehcm, aes(x= Timecode, y=`Respond ID`, fill = storehcm$Channel)   ) +
  geom_bar(colour='black', stat="identity",
           position = position_fill(reverse = TRUE),
           size=.3, inherit.aes = TRUE) +
geom_text(aes(label = round(storehcm$`Respond ID`,2), size = 0.5, hjust = 0.5,vjust = 1),position = position_fill(reverse = TRUE)) 

library(RODBC)
channel <- odbcDriverConnect("Driver={Microsoft Access Driver (*.mdb, *.accdb)};DBQ=G://My Drive//Working//Thuyet trinh//Thuyet trinh//thuyettrinh//Prowtech_Media.accdb")

obdcDataSources("Prowtech_Media.accdb")

data <- sqlQuery( channel , paste ("
SELECT Raw_1_original.TimeCode, Raw_1_original.VehicleTypeCode, Raw_1_original.ProductCode, Raw_1_original.StoreCode, Raw_1_original.Length, Raw_1_original.Seq, Raw_1_original.Position, Raw_1_original.Group
FROM Raw_1_original INNER JOIN Hoang_TimeCode ON Raw_1_original.TimeCode = Hoang_TimeCode.TimeCode"))

#INSERT INTO Raw_1(TimeCode, VehicleTypeCode, ProductCode, StoreCode, Length, Seq, [Position], [Group] )

sqlTables(channel, errors = FALSE, as.is = TRUE,
           catalog = NULL, schema = NULL, tableName = NULL,
           tableType = NULL, literal = FALSE)

save.image()

Step 1: I will need a team to build a connection with Access and SQL to extract data. Benefit: Through Excess, we cannot know what the predecessor did to create table

Step 2: When everything is connected, we can carry on to make them familiar with new connection. Work with .Rproject data then .Accbd

Step 3: Use R language to process table and relationship, in order to give result and put it in Tableau or report template.

Link: https://goo.gl/sz3WiS